BEHAVIORAL GENETICS 97 Understanding the Genetic Basis of Mood Disorders: Where Do We Stand?

Transcription

1 Am. J. Hum. Genet. 60: , 1997 BEHAVIORAL GENETICS 97 Understanding the Genetic Basis of Mood Disorders: Where Do We Stand? Victor I. Reus and Nelson B. Freimer Center for Neurobiology and Psychiatry, Department of Psychiatry, University of California, San Francisco, San Francisco Introduction as the predominant feature of the syndrome. In most current diagnostic systems, mood is viewed as carrying In this issue of the Journal, Sherman et al. (1997) degreater heuristic and practical value than other signs and scribe the promise of genetic approaches for understandsymptoms that might also be present, such as anxiety, ing human behavior and point out a number of obstacles cognitive impairment, or alteration in vegetative functo realization of this promise; these include the methodtions (American Psychiatric Association 1994). Epidemiological challenge of identifying genes for complex traits ological studies have provided strong support for a geand the societal challenge of appropriately using the netic component in the etiology of these disorders information that will be gained if such genetic-mapping (Tsuang and Faraone 1990), particularly in comparison efforts are successful. Genetic-mapping studies in huwith other classes of mental illness. However, as noted mans rest on the premise that traits of interest can be above, our current categorical classification of psychiatreduced to one or more discrete phenotypes and that ric disorders, although reliable, lacks demonstrated etiothese phenotypes result, at least in part, from particular logic validity. It is thus by no means certain that disturalleles at susceptibility loci of reasonably large effect. As bance of mood represents the best indicator of common discussed in this review, abundant evidence suggests that genetic etiology for this class of conditions. For example, severe bipolar mood disorder (BP) fulfills this premise major depressive disorder (MDD), the most common better than other human behavioral traits (Tsuang and mood disorder, is conceptualized as distinct from panic Faraone 1990; Escamilla et al. 1997). The diagnosis of disorder (PD), a common anxiety syndrome, although BP is highly reliable, and its delineation as a distinct two-thirds of patients with PD also will have a lifetime syndrome has proved to be clinically useful in predicting diagnosis of MDD (Breier et al. 1984), and comorbidity course and response to treatment (Goodwin and Jamibetween the two syndromes is common (Kendler et al. son 1990). However, one must keep in mind that this 1987; Coryell et al. 1988). Psychiatric disorders are, diagnostic category, like all psychiatric classifications, is of course, not unique in this regard, and similar issues based on operational criteria (derived from a combinaregarding the etiologic validity of diagnostic constructs tion of epidemiological and clinical observations), rather also complicate genetic mapping of many medical disorthan on any anatomical or physiological evidence. This ders. fact differentiates psychiatric disorders from other etio- The list of mood disorders most commonly includes logically complex categories of disease, such as hyper- MDD, dysthymic disorder (a more chronic but generally tension or diabetes mellitus. In this review we discuss less severe depressive disorder), BP (in which episodes our current understanding of the genetic basis of BP and of major depression alternate with mania), cyclothymic other mood disorders and indicate how our body of disorder (characterized by many episodes of brief and/ knowledge has been influenced by different approaches or less severe depressions alternating with symptoms of to the definition of disease phenotypes. mild mania (hypomania)), and mood disorders due to a The term mood disorders encompasses a group of general medical condition or induced by substance conditions in which a disturbance of mood is recognized abuse. BP is divided further into BP-I (MDD and severe mania) and BP-II (MDD and hypomania), and both of these categories also often are referred to as manic- Received April 15, 1997; accepted for publication April 16, Address for correspondence and reprints: Dr. Victor I. Reus, Center depressive illness. The degree to which these diagnoses for Neurobiology and Psychiatry, Department of Psychiatry, University of California, San Francisco, 401 Parnassus Avenue, San Fran- unknown (Tsuang and Faraone 1990; Pauls et al. 1995; are related genetically or are distinct entities is currently cisco, CA vir@itsa.ucsf.edu Spence et al. 1995). This uncertainty has impeded efforts This article represents the opinion of the authors and has not been to construct realistic models for linkage analysis of BP. peer reviewed by The American Society of Human Genetics. All rights reserved. Furthermore, although most (but not all) genetic-map /97/ $02.00 ping studies treat mood disorders and thought disorders 1283

2 1284 Am. J. Hum. Genet. 60: , 1997 (notably schizophrenia) as independent entities, a sub- were conducted 2 decades ago, using (a) phenotypeassignment stantial number of individuals are afflicted with schizoaffective methods that currently would not be widely disorder, in which disturbances in thought pro- accepted and (b) diagnostic categorizations that differ cesses are as prominent as the mood alteration from those employed in almost all recent genetic-map- (Escamilla et al. 1997). ping studies. An example is the study by Bertelsen et al. (1977), discussed above, which has had a substantial Genetic Epidemiology of Mood Disorders influence on current views regarding the genetic basis of mood disorders. The determination of phenotype in that Uncertainty regarding the etiologic relationship between investigation derived from an unstructured psychiatric BP and other psychiatric disorders has not been interview conducted by a single individual; in contrast, resolved by a large number of family, twin, and adoption most current investigators would accept diagnostic-in- studies aimed at evaluating the degree of familial aggre- terview data only if these were collected by standardized gation and heritability of various mood disorders. Nevertheless, interviews. Additionally, the findings of Bertelsen et al. although family studies of BP, conducted over (1977) were based on diagnostic criteria different than a period of several decades, have utilized several different those currently in use for example, combining individ- designs, including varying definitions of what consti- uals with hypomania and individuals with mania into tutes BP, they consistently have demonstrated familial the same category. Reexamination of epidemiological aggregation of both severe and mild mood disorders. studies such as that by Bertelsen et al. (1977) further One of the most striking findings of these studies has emphasizes the special difficulty of psychiatric genetics; been that BP shows far greater familial aggregation than there is inherent subjectivity in recording even individual other mood disorders. For example, Weissman et al. behavioral features, let alone in determining categorical (1984) observed a relative risk for BP of almost 25, diagnoses that change over time. This fact makes the compared with a relative risk of Ç3.0 for MDD. Similarly, delineation of affected status in current studies as probfor twin studies have indicated a higher heritability lematic as the proper interpretation of past data. BP, compared with that for other forms of mood It is not likely that we will attain a more refined ap- disorder. For example, the classic twin study by Bertel- proach for diagnostic classification of mood disorders son et al. (1977), utilizing the Danish Twin Registry, until we begin to identify the genes that underlie risk indicated a concordance rate, for narrowly defined BP, for these disorders. From that point we may learn to of.79 for MZ twins, versus.19 for DZ twins (as opposed make distinctions finer than are now possible between to.54 and.24, respectively, for major depression). different classes of mood disorders and also may dis- Although most epidemiological studies suggest that cover that particular (and currently unexpected) phenosome cases of MDD probably are genetically related to typic features are etiologically related to one another. BP, only inconclusive results have been obtained from Such use of genetics to redefine phenotypic categories efforts to identify particular MDD subtypes that are has, of course, been one of the most tangible results most likely to reflect a shared genetic etiology with BP of positional cloning of genes responsible for relatively (Goodwin and Jamison 1990). simple disorders. Such diagnostic refinement, however, It is instructive to compare the evidence showing the will not be possible for mood disorders unless the curhigh degree of heritability of narrowly defined BP to rently available classification schemes will permit us to that for either of the traits used for illustrative purposes first map the responsible genes. Genes for several pheno- by Sherman et al. in this issue of the Journal namely, typically complex (nonpsychiatric) disorders have now schizophrenia and emotional stability; for example, the been mapped. Success in mapping such traits usually MZ concordance rate of.79 for BP (Bertelsen et al. has depended on the delineation of narrowly defined 1977) compares with an MZ concordance rate of.46 phenotypes that are most likely to share common genetic for schizophrenia (results from several studies, pooled causation, as determined by examination of the findings by Sherman et al.). The calculated heritability of BP of genetic-epidemiological studies (McInnes and Freimer has been estimated at.59 (Tsuang and Faraone 1990), 1995; Escamilla et al. 1997). These examples could be compared with for emotional stability (as compiled used to guide our approach to genetic-mapping investi- by Sherman et al.). gations of mood disorders; although the epidemiological Sherman et al. point out that twin studies have a number studies of these syndromes are imperfect, they clearly of inherent limitations and that their results there- suggest that severe BP (i.e., BP-I) should be considered fore should be interpreted cautiously. Such caution is such a narrowly defined phenotype. particularly warranted for psychiatric disorders, because of the special complexities that we already have noted. Genetic Mapping Studies of Mood Disorder Relatively little attention has been paid to the fact that The effort to map genes for BP may have attracted the majority of twin studies of psychiatric syndromes more scrutiny from the scientific and general media than

3 Reus and Freimer: Behavioral Genetics has the search for genes responsible for any other human Order Amish community suggest possible SML inheritance trait. Remarkably, several cycles of misplaced excitement only for BP-I, not for other mood disorders (Pauls and exaggeration of failure were widely reported et al. 1995). These studies should not be interpreted as before a single study had even attempted to screen the supporting SML inheritance of BP (Craddock et al. entire genome for linkage to BP. In fact, the initial series 1997; Spence et al. 1997), but they clearly indicate that of studies suggesting localization of BP genes on chromosomes such a model is unlikely for broadly defined mood disor- 11 (among the Old Order Amish) and X ders. Despite these observations, many (but not all) sub- (among non-ashkenazi Israeli Jews) were reported at sequent linkage studies have continued to focus on data a time when the available genetic markers permitted sets in which a high proportion of the genetic information evaluation of linkage across only a tiny fraction of the derives from individuals with diagnoses other than entire genome (Baron et al. 1987; Egeland et al. 1987) BPI. Efforts continue to attempt to identify forms of and when linkage had only been demonstrated for a mood disorder that reflect a common genetic etiology handful of diseases, all with clear-cut Mendelian trans- with BP; these remain unsuccessful (Blacker et al. 1996). mission. These initial linkage studies were predicated Beginning in the early 1990s, the availability of highly on two broad assumptions: first, that relatively simple polymorphic microsatellite markers had a major impact Mendelian models could account for the transmission of on genetic-linkage studies for BP. Several groups rean enormous range of mood-disorder phenotypes and, ported single suggestive localizations for BP genes, loci second, that the power of linkage approaches was sufficiently based on testing multiple markers in several locations great that BP genes could be identified even if but not, however, on complete screening of the genome. the first assumption was incorrect. With the benefit of These studies utilized mainly cosmopolitan collections hindsight it is obvious that such preconceptions were of pedigrees from North America and Europe. Examples naive, although it should be noted that the results of the of these results include possible BP localizations on chro- studies, particularly that of the Amish study, were mosomes 21q (Straub et al. 1994), in the pericentromeric widely accepted until follow-up examination of the same region of 18 (Berrettini et al. 1994), on 16p pedigrees failed to support the original findings (McIn- (Ewald et al. 1995), and on 5p (Kelsoe et al. 1996). The nes and Freimer 1995). It became clear that investigators results of Berrettini et al. (1994) have attracted particu- had failed to account for uncertainty in the assignment lar attention, because they have been supported by an of phenotypes; for example, the fall in the evidence for independent investigation using a different set of North linkage on chromosome 11 in the Amish was due largely American pedigrees (Stine et al. 1995). The results supporting to the development of mood disorder in two previously BP linkage in the pericentromeric region of chroto unaffected individuals whose marker genotypes in this mosome 18 are confined largely to families in which BP chromosome region differed from those of the originally appears to be inherited through the paternal lineage. affected family members (Kelsoe et al. 1989). Addition- This intriguing observation has suggested the possibility ally, the majority of the linkage information in both the of a parent-of-origin effect. It is unclear what mechanism Amish and Israeli studies derived from individuals with would account for such a finding; however, it may be diagnoses other than classic BP for example, unipolar noteworthy that this region contains one of the most MDD (Baron et al. 1987, 1993; Egeland et al. 1987; dramatic instances, in the genome, of suppression of Kelsoe et al. 1989). As indicated above, it already had male recombination relative to female recombination been known, from family studies, that the relative risk (Silverman et al. 1996). for BP was substantially greater than that for MDD Several recently reported linkage studies of BP differ (Weissman et al. 1984) and, from twin studies, that the from previous ones in three important respects: (1) Linkage heritability of BP was substantially greater than that for analysis is focused primarily on individuals with MDD (Bertelsen et al. 1977). severe mood disorder (BP-I only in Ginns et al. 1996; Subsequent to these initial follow-up linkage studies, McInnes et al. 1996; and BP-I and BP-II in Blackwood additional efforts have been made to evaluate the et al. 1996) (2) The bulk of information for linkage degree to which the inheritance patterns of various is derived from affected, rather than from apparently forms of mood disorder are consistent with the models unaffected, individuals (Blackwood et al. 1996; Ginns used in most linkage studies that is, the assumption et al. 1996; McInnes et al. 1996). (3) Results can be that the action of a single major locus (SML) is involved. interpreted in the light of complete screening of the genome Spence et al. (1995), in the largest such study, have in the relevant study populations (Blackwood et shown in a British Columbian sample that, although al. 1996; Ginns et al. 1996; McInnes et al. 1996). there is good evidence for SML inheritance of BP (in What do these studies tell us? In the study by Black- particular, of BP-I), the evidence argues against SML wood et al. (1996), an argument could be advanced for inheritance for MDD and other psychiatric disorders. statistically significant linkage to chromosome 4p. This Even more strikingly, segregation analyses of the Old result was limited to a single extended Scottish pedigree

4 1286 Am. J. Hum. Genet. 60: , 1997 and was not found in other pedigrees from the same population. Additionally, the evidence for linkage depended on the inclusion of BP-II individuals. Ginns et al. (1996), in a new analysis of the Old Order Amish, identified regions on chromosomes 6, 13, and 15 in which excessive allele sharing among affected individuals suggest the possible location of loci for BP-I susceptibility. None of these possible localizations, however, attain statistical significance. McInnes et al. (1996), in a study of two extended pedigrees from the genetically homogenous population of Costa Rica, identified possible localizations for BP-I in several chromosomal regions, notably 11p, 18p, and 18q. The 18q localization was supported by identification of extremely long haplotypes shared by most of the affected individuals in both families, including several alleles that are very rare in the general Costa Rican population (Freimer et al. 1996). In the Costa Rican study the linkage evidence also did not attain statistical significance. Each of these reports, however, identifies regions of the genome that require more intensive study. In two cases (Blackwood et al. 1996; Freimer et al. 1996), the suggested linkage findings are accompanied by extended marker haplotypes in the majority of affected individuals within the relevant pedigrees. These haplotypes indicate that the designated regions are identical by descent (IBD) among affected individuals; since each of these studies draws on subjects from genetically homogeneous populations, there is the possibility of testing whether these regions are shared IBD by other (apparently unrelated) patients from the same populations. This opportunity also offers the hope that the region of IBD sharing can be narrowly delineated on the basis of the expectation that more distantly related affected individuals will share small regions IBD, as a result of recombination over several generations that separate them from their common ancestor(s). Ultimately it may be possible, by use of such isolated populations, to define, for BP loci, candidate regions sufficiently narrow to permit positional cloning of causative genes. It is unlikely that a single major susceptibility gene exists that is responsible for a high proportion of BP cases in all populations. How, then, should we interpret the fact that several genome-screening studies have highlighted multiple possible chromosomal regions that could contain BP-susceptibility genes? Genetic-mapping studies of complex traits may be viewed as a highly iterative process, in which initial genome-screening studies serve to propose hypotheses to be tested through additional investigation (McInnes et al. 1996). According to this view, one does not expect the explosive progress toward gene identification that has characterized the study of simple Mendelian traits (Jamison and McInnis 1996), and we can anticipate that many (if not most) of the currently suggested BP localizations will prove spurious. In this respect the results of genomescreening studies of BP are roughly comparable to those recently observed for other complex traits, such as insulin-dependent diabetes mellitus and multiple sclerosis (Davies et al. 1994; Hashimoto et al. 1994; Ebers et al. 1996; Haines et al. 1996; Kuokkanen et al. 1996; Sawcer et al. 1996). Some commentators have suggested that the difficulties and uneven rate of progress inherent in mapping genes for complex traits are somehow particular to BP, perhaps because of the glib references that can be made to the euphoria of apparent success and the dysphoria of apparent failure (Morell 1996; Risch and Botstein 1996). Such publicity may have the unfortunate result of generating undue pessimism, among the scientific community and the general public, regarding the possibility of ultimately elucidating the genetic basis of BP and other mood disorders. Societal Implications Even if, in the near future, one or more susceptibility genes for bipolar disorder are identified, it will be difficult, as Sherman et al. point out, to know how to utilize such information clinically. The societal stigma of bipo- lar disorder and other psychiatric conditions, as well as their variable course, make it likely that the problems already identified in the genetic testing of other common complex diseases, such as breast cancer and colon can- cer, will be multiplied when genetic testing in psychiatric disorders becomes feasible. The opportunity exists, however, to identify the salient issues and to plan for the day when genetic risk of psychiatric illness can be assessed reliably. For example, our group has surveyed individuals with mood disorder, their friends and rela- tives, and health-care providers, to explore the implications of genetic testing in bipolar disorder (Smith et al. 1996). We have observed that reproductive decisions and attitudes toward presymptomatic testing in children differ markedly between consumer groups and healthcare providers and that they depend greatly on presumed severity of illness and the availability of as yet undetermined preventive strategies that might ameliorate the severity or course of illness. Individuals who are likely to pursue genetic testing may find it difficult to accept the inherent uncertainty in the prediction of genetic risk and may misinterpret the information in such a way as to adversely affect their own and/or other family member s life experiences. In our study, psychiatric physi- cians were much more likely to endorse termination of a pregnancy than were either patients and support-group members or medical students, if informed of a probable fetal inheritance of bipolar disorder. This difference in interpretation of data was particularly evident in scenarios in which the likelihood of developing BP disorder was high and in which the course of illness was severe.

5 Reus and Freimer: Behavioral Genetics Surprisingly, patients and support-group members exlinked CR, Hostetter AM, et al (1987) Bipolar affective disorders pressed greater interest in genetic testing than did psychiatrists, to DNA markers on chromosome 11. Nature 325: even when it was assumed that no prophylactic treatment existed. The need for further study of attitudes Escamilla MA, Freimer NB, Reus VI (1997) The genetics of bipolar disorder and schizophrenia. In: Rosenberg RN, toward mood disorders is clear. It most likely will be at Prusiner SB, DiMauro S, Barchi RL (eds) The molecular and least a few years before genes for BP are identified. This genetic basis of neurological disease, 2d ed. Butterworthprovides us an opportunity to investigate systematically Heinemann, Boston, pp the societal implications of the ability to determine ge- Ewald H, Mors O, Flint T, Koed K, Eiberg H, Kruse TA (1995) netic risk for this disorder before the reality is thrust A possible locus for manic depressive illness on chromosome upon us. 16p13. Psychiatr Genet 5:71 81 Freimer NB, Reus VI, Escamilla MA, McInnes LA, Spesny M, Leon P, Service SK, et al (1996) Genetic mapping using References haplotype, association and linkage methods suggests a locus for severe bipolar disorder (BPI) at 18q22-q23. Nat Genet American Psychiatric Association (1994) Diagnostic and statistical 12: manual of mental disorders, 4th ed. American Psychi- Ginns EI, Ott J, Egeland JA, Allen CR, Fann CS, Pauls DL, atric, Washington Weissenbach J, et al (1996) A genome-wide search for chro- Baron M, Freimer N, Risch N, Lerer B, Alexander JR, Straub mosomal loci linked to bipolar affective disorder in the Old RE, Asokan S, et al (1993) Diminished support for linkage Order Amish. Nat Genet 12: between manic depressive illness and X-chromosome mark- Goodwin FK, Jamison KR (1990) Manic-depressive illness. ers in three Israeli pedigrees. Nat Genet 3:49 55 Oxford University Press, New York Baron M, Risch N, Hamburger R, Mandel B, Kushner S, New- Haines JL, Ter-Minassian M, Bazyk A, Gusella JF, Kim DJ, man M, Drumer D, et al (1987) Genetic linkage between Terwedow H, Pericak-Vance MA, et al (1996) A complete X-chromosome markers and bipolar affective illness. Nature genomic screen for multiple sclerosis underscores a role for 326: the major histocompatibility complex. Nat Genet 13:469 Berrettini WH, Ferraro TN, Goldin LR, Weeks DE, Detera- 471 Wadleigh S, Nurnberger JI Jr, Gershon ES (1994) Chromosome Hashimoto L, Habita C, Beressi JP, Delepine M, Besse C, 18 markers and manic-depressive illness: evidence for Cambon-Thomsen A, Deschamps I, et al (1994) Genetic a susceptibility gene. Proc Natl Acad Sci USA 91:5918 mapping of a susceptibility locus for insulin-dependent dia betes mellitus on chromosome 11q. Nature 37: Bertelsen A, Harvald B, Hauge M (1977) A Danish twin study Jamison KR, McInnis MG (1996) Genetic studies of manicof manic-depressive disorders. Br J Psychiatry 130: depressive illness. Nat Med 2: Blacker D, Faraone SV, Rosen AE, Guroff JJ, Adams P, Weissman Kelsoe JR, Ginns EI, Egeland JA, Gerhard DS, Goldstein AM, MM, Gershon ES (1996) Unipolar relatives in bipolar Bale SJ, Pauls DL, et al (1989) Re-evaluation of the linkage pedigrees: a search for elusive indicators of underlying bipolarity. relationship between chromosome 11p loci and the gene for Am J Med Genet 67: bipolar affective disorder in the Old Order Amish. Nature Blackwood DH, He L, Morris SW, McLean A, Whitton C, 342: Thomson M, Walker MT, et al (1996) A locus for bipolar Kelsoe JR, Sadovnick AD, Kristbjarnarson H, Bergesch P, affective disorder on chromosome 4p. Nat Genet 12:427 Mroczkowski-Parker Z, Drennan M, Rapaport MH, et al 430 (1996) Possible locus for bipolar disorder near the dopamine Breier A, Charney DS, Heninger GR (1984) Major depression transporter on chromosome 5. Am J Med Genet 67:533 in patients with agoraphobia and panic disorder. Arch Gen 540 Psychiatry 41: Kendler KS, Heath AC, Martin NG, Eaves LJ (1987) Symp- Coryell W, Endicott J, Andreasen NC, Keller MB, Clayton PJ, toms of anxiety and symptoms of depression: same genes, Hirschfeld RM, Scheftner WA, et al (1988) Depression and different environments? Arch Gen Psychiatry 44: panic attacks: the significance of overlap as reflected in fol- Kuokkanen S, Sundvall M, Terwilliger JD, Tienari PJ, Wikstrom low-up and family study data. Am J Psychiatry 145:293 J, Holmdahl R, Pettersson U, et al (1996) A putative 300 vulnerability locus to multiple sclerosis maps to 5p14-p12 Craddock N, VanEerdewegh P, Reich T (1997) Single major in a region syntenic to the murine locus Eae2. Nat Genet locus models for bipolar disorder are implausible. Am J Med 13: Genet 74:18 McInnes LA, Escamilla MA, Service SK, Reus VI, Leon P, Silva Davies JL, Kawaguchi Y, Bennett ST, Copeman JB, Cordell S, Rojas E, et al (1996) A complete genome screen for genes HJ, Pritchard LE, Reed PW, et al (1994) A genome-wide predisposing to severe bipolar disorder in two Costa Rican search for human type 1 diabetes susceptibility genes. Na- pedigrees. Proc Natl Acad Sci USA 23: ture 37: McInnes LA, Freimer NB (1995) Mapping genes for psychiat- Ebers GC, Kukay K, Bulman DE, Sadovnick AD, Rice G, ric disorders and behavioral traits. Curr Opin Genet Dev 5: Anderson C, Armstrong H, et al (1996) A full genome search in multiple sclerosis. Nat Genet 13: Morell V (1996) Manic-depression findings spark polarized Egeland JA, Gerhard DS, Pauls DL, Sussex JN, Kidd KK, Allen debate. Science 272:31 32

6 1288 Am. J. Hum. Genet. 60: , 1997 Pauls DL, Bailey JN, Carter AS, Allen CR, Egeland JA (1995) Spence MA, Flodman PL, Sadovnik AD, Ameli H, Remick RA Complex segregation analyses of old order Amish families (1997) Single major locus models for bipolar disorder are ascertained through bipolar I individuals. Am J Med Genet implausible response. Am J Med Genet 74: : Spence MA, Flodman PL, Sadovnick AD, Bailey-Wilson JE, Risch N, Botstein D (1996) A manic depressive history. Nat Ameli H, Remick RA (1995) Bipolar disorder: evidence for Genet 12: a major locus. Am J Med Genet 60: Sawcer S, Jones HB, Feakes R, Gray J, Smaldon N, Chataway Stine OC, Xu J, Koskela R, McMahon FJ, Gschwend M, Friddle J, Robertson N, et al (1996) A genome screen in multiple C, Clark CD, et al (1995) Evidence for linkage of bipolar sclerosis reveals susceptibility loci on chromosome 6p21 and disorder to chromosome 18 with a parent-of-origin effect. 17q22. Nat Genet 13: Am J Hum Genet 57: Sherman SL, DeFries JC, Gottesman II, Loehlin JC, Meyer JM, Straub RE, Lehner T, Luo Y, Loth JE, Shao W, Sharpe L, Pelias MZ, Rice J, et al (1997) Recent developments in human Alexander JR, et al (1994) A possible vulnerability locus behavioral genetics: past accomplishments and future direc- for bipolar affective disorder on chromosome 21q22.3. Nat tions. Am J Hum Genet 60: (in this issue) Genet 8: Silverman GA, Overhauser J, Gerken S, Aburomia R, O Con- Tsuang MT, Faraone SV (1990) The genetics of mood disorders. nell P, Krauter KS, Detera-Wadleigh SD, et al (1996) Report Johns Hopkins University Press, Baltimore of the Fourth International Workshop on Human Chromo- Weissman MM, Gershon ES, Kidd KK, Prusoff BA, Leckman some 18 Mapping: Boston, Massachusetts, October 7 9, JF, Dibble E, Hamovit J, et al (1984) Psychiatric Cytogenet Cell Genet 75: disorders in the relatives of probands with affective disorders: Smith LB, Sapers B, Reus VI, Freimer NB (1996) Attitudes the Yale University National Institute of Mental towards bipolar disorder and predictive genetic testing Health Collaborative Study. Arch Gen Psychiatry 41: among patients and providers. J Med Genet 33: